This week has been very eventful for me. For one thing CMS had a very successful global run where we collected so many cosmic ray events that we decided to give ourselves a weekend off, but at the same time we were also doing a Computing Challenge (which Ken can surely write about) to test that we can ship the LHC data around as quickly as it arrives (and do some basic operations on it). For now we were using fake (so simulated) data for the latter, as even though the cosmic run was successful, the data volume was no way near what we expect once we get collisions.
So how did those activities affect me? Well, I am responsible for the databases and software that ensures we can read out the CMS pixel detector. This meant I had to check that the pixel data was coming out of the cosmic run (it was) and that the simulated pixel data was correct so the the people responsible for analyzing the data can check that their software works. Which effectively meant being the first person to look at all data. Of course simulated data is something we have been dealing with for the last few years, so at least from the pixel side there was not much excitement there. However, I was really looking forward to looking at the first *real* hits coming from a pixel detector in the real system.
To give you some ideas of things that you have to deal with in these matters let me explain the configuration in the cosmic run. As the CMS pixel detector can only be installed once the CMS beam pipe (the part of the accelerator that goes through the CMS detector) is installed, we instead connected only a small part of the pixel detector to the readout electronics. This small part we call the Pixel In a Box (PIB, in CMS you need to have an acronym for your activity. It almost seems you do not get taken seriously otherwise). In a few weeks we will have a real detector installed, it is ready and we’re keeping it at a random cleanroom at CERN for now (and the other half of it in Zuerich) but the beam pipe needs to be ready first. The PIB was connected to the entire system just like the real pixel detector would be, and taking data just like the rest of the detector. Of course a small pixel detector (it’s about 10 square inches of active material in total) does not find many cosmic rays so this really was just a exercise in cooperating with the rest of the detector. Still, exciting and definitely a new phase.
The pixel in a box run was a success, data was taken and it very quickly arrived at the lowest level grid analysis facilities (the CERN tier zero) where the responsible expert (read: me) took a look at it. And then the trouble started. There was nothing in the data! Of course the first step is to prove I wasn’t doing something wrong myself, but a few days and many discussions with the other experts later it turned out that…. the PIB was supposed to create no data. Actually the PIB was set up in such a way that all pixels were turned off. In some sense it is a good thing that I could prove this but some way or another it just does not feel the same as when you get actual data out.
I now try to comfort myself by telling myself that at least I proved I could see there was no data before someone standing next to the PIB could see there was no data. But it still is a bit disappointing.
Tags: CMS
